Skip to content

ENH: broaden SMILES handling to match RDKit #60

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
tylerjereddy wants to merge 2 commits into
base: master
Choose a base branch
from
+23 −1
Open

ENH: broaden SMILES handling to match RDKit #60

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
cc6ce6a
ENH: broaden SMILES handling to match RDKit
tylerjereddy Jun 26, 2026
45e6fd9
MAINT: PR 60 revisions
tylerjereddy Jun 26, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 6 additions & 1 deletion pysmiles/read_smiles.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,6 +18,7 @@
"""

import enum
import re
import logging

import networkx as nx
Expand Down Expand Up @@ -244,10 +245,14 @@ def read_smiles(smiles, explicit_hydrogen=False, zero_order_bonds=True,
information.
Edges will have an 'order'.
"""
# sanitize invalid SMILES input that is accepted by RDKit
# see: https://github.com/gruenewald-lab/CGsmiles/issues/70#issuecomment-4750353505
pattern = r'(\(=[A-Z]\))(\d)'
mod_smiles = re.sub(pattern, r'\2\1', smiles)

@tylerjereddy tylerjereddy Jun 26, 2026

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I suppose one could argue that this is the beginning of a slippery slope to shim around more and more pathological SMILES (RDKit can permute at least a dozen variants of the SMILES inputs used in this PR and still handles them just fine).

It might be useful to clarify why we're not using RDKit in this ecosystem--just the "weight" of the dependency?

bond_to_order = {'-': 1, '=': 2, '#': 3, '$': 4, ':': 1.5, '.': 0}
default_bond = 1
default_aromatic_bond = 1.5
mol, ez_isomer_atoms, ring_bonds = base_smiles_parser(smiles, strict=strict,
mol, ez_isomer_atoms, ring_bonds = base_smiles_parser(mod_smiles, strict=strict,
node_attr='_atom_str', edge_attr='_bond_str')
for node in mol:
mol.nodes[node].update(parse_atom(mol.nodes[node]['_atom_str']))
Expand Down
22 changes: 22 additions & 0 deletions tests/test_read_smiles.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1022,3 +1022,25 @@ def test_aromatic_molecules(smiles):
"""These molecules are totally aromatic"""
mol = read_smiles(smiles, reinterpret_aromatic=True)
assert all(nx.get_node_attributes(mol, 'aromatic').values())


@pytest.mark.parametrize("smiles", [
# these are both interpreted as the same
# molecule in RDKit
"C(c1c2cccc3c2c(cc1)C(=C)C(=C)C(=C)3)",

@tylerjereddy tylerjereddy Jun 26, 2026

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The other pathological example from the cross-listed issue is far harder to shim around. My one hesitation is the danger of reinventing the wheel given the versatility RDKit has with SMILES canonicalization/permutations.

"C(c1c2cccc3c2c(cc1)C(=C)C(=C)C3(=C))"

@tylerjereddy tylerjereddy Jun 26, 2026

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

for reference:

from rdkit import Chem
smiles_1 = "C(c1c2cccc3c2c(cc1)C(=C)C(=C)C(=C)3)" 
smiles_2 = "C(c1c2cccc3c2c(cc1)C(=C)C(=C)C3(=C))" 
mol_1 = Chem.MolFromSmiles(smiles_1)
mol_2 = Chem.MolFromSmiles(smiles_2)
assert Chem.MolToSmiles(mol_1) == Chem.MolToSmiles(mol_2)

and neither of those original SMILES matches what RDKit considers "canonical." I'm assuming the opensmiles specification is therefore different from what they consider standard/canonical.

])
def test_non_canonical_smiles_handling(smiles):
# harmonizing SMILES input handling to better match RDKit
# can help accommodate SMILES strings provided by chemists
# per:
# https://github.com/gruenewald-lab/CGsmiles/issues/70
mol = read_smiles(smiles)
# expected values are from the "good" string at:
# https://github.com/gruenewald-lab/CGsmiles/issues/70#issuecomment-4750353505
expected_nodes = list(range(17))
expected_edges = [(0, 1), (1, 2), (1, 10), (2, 3), (2, 7), (3, 4), (4, 5),
(5, 6), (6, 7), (6, 15), (7, 8), (8, 9), (8, 11), (9, 10),
(11, 12), (11, 13), (13, 14), (13, 15), (15, 16)]
assert list(mol.nodes) == expected_nodes
assert list(mol.edges) == expected_edges
Loading